[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CA+TsHUC6+xEW4ZB5mBHMf=hdnjPBQaruAGhdEE=p+ad5tRkRJQ@mail.gmail.com>
Date: Fri, 23 Aug 2013 10:13:45 +0530
From: Sayantan Datta <std2048@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: *pcount type should be uint64_t
On Fri, Aug 23, 2013 at 1:33 AM, Solar Designer <solar@...nwall.com> wrote:
> You may calculate how large *pcount would be, and if it'd exceed a
> certain threshold (which you'd keep below 2G), then reduce the portion
> of mask processed on GPU until the would be *pcount value is sane.
>
tuning the mask processed inside GPU would be very coarse. E.g If say two
mask placeholders are processed inside gpu, then lowering portion of mask
processed on gpu would mean rejecting an entire placeholder. For a ?l
placeholder it would mean lowering keys processed inside gpu by 26 times,
which is unacceptable from performance perspective. I don't think general
user will bother to provide very specific mask that would minimize kernel
run time as well as maintain good performance.
I think in that respect kpc is more tunable. I'll minimize kpc as low as I
can without impacting too much say less than 10% of peak performance.
Regards,
Sayantan
Content of type "text/html" skipped
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
